Qoder CN Series:Best practices for enterprise code completion enhancement

Guidelines for selecting and preparing code repositories

The following sections describe how to prepare code repositories in two different scenarios:

• Scenario 1: Daily engineering development — improve R&D efficiency. In enterprise development, you can use multiple code repositories to support retrieval augmentation, which improves code reuse and development efficiency. The following are the main ways to select code repositories:

• Scenario 2: Specific business scenarios — ensure code consistency and reduce repeated development. In enterprise application development, some business logic must stay consistent to ensure system stability and maintainability. Without unified standards, different developers may adopt different implementations, which leads to inconsistent business logic, more complex code maintenance, and potential system stability issues.

/**
 * Use the enterprise-developed distributed lock framework for concurrency control
 */
@Service
public class InventoryService {

    @Autowired
    private DistributedLockFramework lockFramework;

    @Autowired
    private InventoryRepository inventoryRepository;

    /**
     * Deduct inventory. Use a distributed lock to ensure concurrency safety.
     * @param productId The product ID.
     * @param quantity The quantity to deduct.
     * @return Whether the deduction is successful.
     */
    public boolean deductInventory(Long productId, int quantity) {
        String lockKey = "inventory:" + productId;
        return lockFramework.withLock(lockKey, 30, TimeUnit.SECONDS, () -> {
            Inventory inventory = inventoryRepository.findByProductId(productId);
            if (inventory == null || inventory.getQuantity() < quantity) {
                return false;
            }
            inventory.setQuantity(inventory.getQuantity() - quantity);
            inventoryRepository.save(inventory);
            return true;
        });
    }
}

/**
 * Recommendation engine service
 * Implements a personalized recommendation mechanism based on user behavior and commodity features
 */
@Service
public class RecommendationService {

    @Autowired
    private UserBehaviorRepository userBehaviorRepo;

    @Autowired
    private ProductRepository productRepo;

    /**
     * Generate a personalized recommendation list for a specified user
     * @param userId The user ID.
     * @param limit The number of recommendations.
     * @return The list of recommended commodity IDs.
     */
    public List<Long> generateRecommendations(Long userId, int limit) {
        // Obtain user behavior data
        List<UserBehavior> behaviors = userBehaviorRepo.findRecentByUserId(userId, 100);

        // Extract user interest features
        Map<String, Double> userInterests = extractUserInterests(behaviors);

        // Obtain the candidate commodity pool
        List<Product> candidateProducts = productRepo.findRecentlyActive(1000);

        // Score and rank commodities
        List<ScoredProduct> scoredProducts = candidateProducts.stream()
            .map(product -> new ScoredProduct(product, calculateScore(product, userInterests)))
            .sorted(Comparator.comparing(ScoredProduct::getScore).reversed())
            .collect(Collectors.toList());

        // Return the highest-scoring commodities
        return scoredProducts.stream()
            .limit(limit)
            .map(sp -> sp.getProduct().getId())
            .collect(Collectors.toList());
    }

    // Other auxiliary methods...
}

Code file specifications

• Supported languages and frameworks:

  • Backend: Java, C#, C/C++, Go, Python.

  • Frontend: JavaScript, TypeScript, Vue, React.

• Upload limits: Only source code files are accepted. The code repository should contain only the source code files that are actually written. For example, upload .java files for Java, .cs files for C#, and .js or .jsx files for JavaScript.

• Avoid uploading the following content:

  • Test data and code: Do not upload test scripts, test cases, or any test-related code that does not contain business logic.

  • Mock methods: Exclude all code generated by mock methods and tools, unless the code contains specific implementations of business logic.

  • Build artifacts:

    • Frontend: Exclude files generated by build tools such as Webpack and Gulp. These files are usually located in the dist or build directory.

    • Backend: Exclude compiled DLL files and all other compilation output.

• Comment requirements:

  • Add detailed comments at the header of each function that you want to be retrieved.

  • Comments should provide enough information to distinguish different functions. Reference a comment template or adjust comments according to your enterprise specifications.

/**
 * Update the status of a specified order.
 *
 * @param orderId The unique identifier of the order.
 * @param newStatus The new order status.
 * @return boolean Indicates whether the update is successful.
 */

• Function naming requirements:

  • If function comments are brief, the function name must accurately describe its function.

  • Use clear and descriptive names, such as exportOrdersToPDF and updateOrderStatus, instead of func1.

Upload guidelines

• Package compressed files: Package code files into .zip, .gz, or .tar.gz format.

• Code package size limit: Each code package must not exceed 100 MB.

Backend best practices

Generate code from natural language comments

• Upload code to the enterprise code repository: Upload a compressed package that contains the required function code to the enterprise code repository, for example the Snowflake algorithm code, and make sure that the target function follows the comment specification with comments placed at the function header. For more information about preparing the code repository, see How administrators prepare a high-quality enterprise code repository.

/**
 * Use the Snowflake algorithm to generate a unique serial number
 * @param workerId
 * @return
*/
public synchronized Long getSnowFlowerId(long workerId){
 long id = -1L;

 if (workerId < 0 || workerId > snowFlowerProperties.getMaxWorkerId()) {
    throw new IllegalArgumentException(
      String.valueOf("workerID must gte 0 and lte " + snowFlowerProperties.getMaxWorkerId()));
 }

 // ... algorithm implementation code ...

return id;
}

• Enter comments: In the IDE, locate a Java class and enter a comment that matches the function you want to recall. The comment format can be flexible, but make sure the meaning is accurate and consistent.

//Please generate code that uses the Snowflake algorithm to generate a unique ID and returns the generated ID

/**
 * Use the Snowflake algorithm to generate a unique serial number
 * @param wId
 * @return
*/

Comment requirements:

• Comment length: Avoid overly short comments when you write code. Use at least 15 characters. Comments that are too short cannot trigger a recall.

• Comment semantics: Make sure the semantics of comments are accurate and meaningful. Include keywords and return value descriptions where possible, so that Qoder CNcan accurately understand and match the corresponding code.

• Multi-language support: Both Chinese and English comments are supported. The comments in the code repository and the comments used during actual coding can be in different languages.

• Parameter name flexibility: Parameter names can be handled flexibly. Qoder CNautomatically adjusts to the provided parameters to match the recalled code. The following are negative examples:

  • //Snowflake algorithm — Issue: Not enough information is provided, and the comment is too short.

  • //Generate a unique serial number — Issue: No specific keywords are used, which may affect understanding and matching.

• Code generation: After you press Enter the first time, Qoder CNprovides a completion suggestion generated from the comment. After you press Enter again, Qoder CNcompletes the code based on the code in the enterprise code repository.

/**
 * Use the Snowflake algorithm to generate a unique serial number
 * @param workerId
 * @return
 */
public synchronized Long get(long workerId) {
    long id = -1L;
    if (workerId < 0 || workerId > snowFlowerProperties.getMaxWorkerId()) {
        throw new IllegalArgumentException(
                String.valueOf("workerID must gte 0 and lte " + snowFlowerProperties.getMaxWorkerId()));
    }
    long timestamp = timeGen();
    if (timestamp < lastTimestamp) {
        long refusedSeconds = (lastTimestamp - timestamp) / 1000;
        throw new RuntimeException(
                String.valueOf("Clock moved backwards. Refusing for " + refusedSeconds + " seconds"));
    }
    if (lastTimestamp == timestamp) {
        // Within the same millisecond, the sequence number is incremented
        sequence = (sequence + 1) & snowFlowerProperties.getSequenceMask();
        if (sequence == 0) {
            // The maximum number of sequence numbers in the same millisecond has been reached
            timestamp = tilNextMillis(lastTimestamp);
        }
    } else {
        // In different milliseconds, the sequence number is set to 0
        sequence = 0L;
    }
    lastTimestamp = timestamp;
    // Timestamp part
    id = (timestamp - snowFlowerProperties.getTwepoch()) << snowFlowerProperties.getTimestampLeftShift();
    // Worker identifier part
    id |= (workerId << snowFlowerProperties.getWorkerIdShift());
    // Sequence number part
    id |= sequence;
    return id;
}

• If your comment contains parameters, Qoder CNautomatically adjusts the parameters in the generated code to ensure naming consistency.

• To refresh the cache and get new completion suggestions, press ⌥(option) + P on macOS or Alt + P on Windows to manually trigger inline completion.

Generate code from function signatures

• Upload code to the code repository: Upload a compressed package that contains the required function code to the enterprise code repository, and make sure these functions have clear and unique identifiers for retrieval and recognition. For more information about preparing the code repository, see How administrators prepare a high-quality enterprise code repository.

• Enter function signatures: In the IDE, locate a Java class and enter the signature of the target function. Parameter names can be handled flexibly. Qoder CNautomatically adjusts to the provided parameters to match the recalled code.

public List<Object> nextList(String name, int size)

Function signature requirements:

• Function name: Use a clear function name with enough semantics to serve as the basis for similarity matching.

• Parameters and return values: The types and order must match the target function, but parameter names can be handled flexibly. Qoder CNautomatically adjusts to the provided parameters to match the recalled code. The following are negative examples:

  • public List<Object> func1(String name, int size) — Issue: The function name has unclear semantics and cannot accurately reflect the function.

  • public List<String> nextList(int orderId) — Issue: The parameter types and return value type do not match the target function.

• Completion suggestion: After you press Enter the first time, Qoder CNprovides a code completion suggestion. After you press Enter again, Qoder CNautomatically completes the code based on the code in the enterprise code repository.

• Qoder CNautomatically adjusts the parameter names in the generated code based on the parameter names you provide, to ensure naming consistency.

• To refresh the cache and get new completion suggestions, press ⌥(option) + P on macOS or Alt + P on Windows to manually trigger inline completion.

Frontend best practices

Complete code of frontend custom components from tags

• Upload code to the code repository: Before you start, make sure all required frontend component code is uploaded to the enterprise code repository. The following is an example of the React framework:

<LTable
      isReady={isReady}
      formInitialValues={formInitialValues}
      rowKey="key"
      tableRef={tableRef}
      toolbarLeft={
          <Button type="primary">Add</Button>
      }
      formItems={formItems}
      formRef={formRef}
      columns={columns}
      request={async (params, requestType) => {
        const res: Record<string, any> = await apiGetUserList(params);
        return {
          data: res.data,
          total: res.total,
        };
      }}
/>

• Write component code: Open the corresponding .jsx file in your IDE and start writing code. Enter a basic HTML tag or a custom component tag, such as <LTable />.

• Automatic code completion: When the code you enter reaches a certain length and matches code in the enterprise component library, the IDE automatically triggers code completion and generates the complete component code. You can also press Enter to trigger code completion manually.

Trigger completion within a complete component tag.

Generate code from natural language comments

• Upload code to the code repository: Upload a compressed package that contains the required function code to the enterprise code repository, and make sure each function follows the comment specification with comments placed at the function header. For more information about preparing the code repository, see How administrators prepare a high-quality enterprise code repository. The following is a JavaScript example:

/**
 * Generate an object keyed by id based on error messages
 * @param {Array<validator,Result>} results
 * @return {Record<string,string>}
*/
function getErrObj(results) {
  // ... function implementation code ...
}

• Enter comments: In the IDE, enter specific comment content in a JavaScript file, as shown in the following example:

//Generate an object keyed by id based on error messages

Comment requirements:

• Comment length: Avoid overly short comments when you write code. Use at least 15 characters. Comments that are too short cannot trigger a recall.

• Comment semantics: Make sure the semantics of comments are accurate and meaningful. Include keywords and return value descriptions where possible, so that Qoder CN can accurately understand and match the corresponding code.

• Multi-language support: Both Chinese and English comments are supported. The comments in the code repository and the comments used during actual coding can be in different languages.

• Parameter name flexibility: Parameter names can be handled flexibly. Qoder CNautomatically adjusts to the provided parameters to match the recalled code.

• Code generation: After you press Enter the first time, Qoder CNprovides a completion suggestion generated from the comment. After you press Enter again, Qoder CNcompletes the code based on the code in the enterprise code repository.

//Generate an object keyed by id based on error messages
export function getErrObj(results) {
    if (!results || results.length <= 0) return {};

    const obj = {};
    for (let i = 0; i < results.length; i++) {
        const result = results[i];
        if (result.ok === false) {
            obj[result.id] = result.msg || '';
        }
    }
    return obj;
}

• If your comment contains parameters, Qoder CNautomatically adjusts the parameter names in the generated code to ensure naming consistency.

• To refresh the cache and get new completion suggestions, press ⌥(option) + P on macOS or Alt + P on Windows to manually trigger inline completion.

FAQ

After reinstallation, code in the knowledge base still cannot be recalled, even after you restart the IDE or log on again.

Solution:

• On macOS, run the following command to restart the process and clear the cache:

  ps -ef|grep lingma|grep start|awk '{print $2}'|xargs -I {} kill -9 {}

• On Windows, end the Qoder CN process in Task Manager.